What the Ansoff matrix is
The matrix is a two-by-two grid. One axis is the product, meaning the offer the business sells: existing or new. The other is the market, meaning the group of buyers it sells to: existing or new. A market here can be a segment, a region, a country, a customer size or a sales channel, as long as it is a group of buyers the business does not currently reach. Every growth idea can be placed in one of the four cells, and the cell says what kind of move it is.
The tool earns its place at the moment a team asks where next year's growth will come from. That question usually arrives after a stall, as on the page on marketing growth that has plateaued, or during annual planning, and it tends to be answered by whichever idea is most exciting in the room. The matrix slows that down. It asks for at least one option in each direction and then asks how much is known about each, so that a new product for a new audience is compared with the plainer options before it is chosen.
What the matrix makes visible is newness, and newness is a proxy for uncertainty. In the existing-market, existing-product cell the team knows the buyers, the messages that work, the channels and the cost of a sale. Each step away from that cell replaces some of that knowledge with assumptions. What the matrix hides is just as important: it says nothing about how large or profitable an option is, whether the business can deliver it, or in what order options should be tried. Those answers come from the guides and frameworks it hands over to.
The four directions
| Move | What is new | Typical marketing work | Main risk |
|---|---|---|---|
| Market penetration | Nothing: the same product sold to the same kind of buyer | Winning share from rivals, raising purchase frequency, cross-selling the range, reducing customer losses, pricing and promotion | Saturation: the reachable buyers run out, each extra sale costs more and price competition erodes margin |
| Market development | The buyers: a new segment, region, country or channel | Researching the new buyers, adapting the message and channels, localising, building proof for an audience that has never heard of the business | The product's appeal or the channel economics may not transfer, and the team does not yet know these buyers |
| Product development | The offer: a new product, line, tier or service for current buyers | Concept testing with existing customers, launching to the base, bundling and cross-selling the new item | Buyers may not want it, it may take sales from the core range, and development cost arrives before any revenue |
| Diversification | Both: a new product for buyers the business does not serve | A full launch with new research, positioning, channels and proof | Two sets of unknowns at once, with little of the existing brand, data or know-how carrying across |
Market penetration is the closest to the current plan, and much of a mature team's work sits here: better conversion, more frequent purchase, keeping more customers, and Cross-SellingOffering a customer a related product alongside the one they are buying; upselling offers a higher-value version instead. to buyers who already trust the business. A falling Churn RateShare of customers at the start of a period who left during it. counts as penetration growth too, because the business holds more of a market it already serves. The limit is the size of that market. When most of the reachable buyers already buy, the only remaining levers are price and share taken from rivals, both of which get expensive.
Market development keeps the product and changes the buyer. It depends on SegmentationDividing a market or customer list into groups that need different offers or messages. work, because the first question is which new group has a real need for the existing offer. A new segment at home, a new country and a new sales channel all sit in this cell, though they carry very different amounts of unknown.
Product development keeps the buyer and changes the offer. Its advantage is that the business already knows whom it is selling to and can ask them before building anything. Its risk is that a new product competes for the same customers' budget as the old one.
Diversification changes both, which is why it is the move most often chosen for its appeal and least often for its evidence. It is not forbidden; some businesses find their best growth there. It simply needs the most research and the most patience, and it should be compared against the other three before it wins.
Applying it: listing options and gathering evidence for each cell
Start by defining the current product and market precisely, because the cells depend on that definition. "We sell coffee online" makes almost everything look like penetration; "we sell whole-bean coffee subscriptions to home brewers in one country through our own website" makes the boundaries clear, and shows that a new channel or a new pack format is a real move.
Then fill each cell with at least one candidate. Good sources are customer requests and complaints, lost-sale notes from sales, the products customers buy elsewhere alongside yours, the segments that already find you without any marketing, and the moves competitors have made. An empty cell is a finding in itself: it usually means the team has not looked, not that no option exists.
For each candidate, write down three things: the evidence that demand exists, the evidence that the business can reach and serve those buyers, and the cost of the smallest test that would replace an assumption with a fact. A penetration option might be testable within a month using current channels. A diversification option might need research, a prototype and a new channel before the first real signal arrives. Laying those side by side is the point of the exercise, because it shows what each direction would cost to learn about, not only what it might be worth.
The shortlist then moves to other tools. The market sizing guide estimates how large the serviceable and obtainable market is for each surviving option, which often removes an attractive idea that turns out to be small. For a move into new countries, the guide to market entry and prioritisation scores candidate markets and sets their order. Once one move is chosen, the go-to-market planning guide turns it into a launch with a first segment, an offer, channels and phase gates, and the Go-to-Market StrategyThe plan for how a product or offer reaches and wins its target customers. plan is where the matrix's job ends.
Worked example (illustrative)
A coffee roaster lists one growth option per cell
Kettlewhin Coffee Roasters, an invented speciality roaster, sells beans online and to 40 independent cafés. Revenue has been flat at about £1.2 million for two years. The team lists one option per cell. Penetration: grow subscriptions among existing online buyers, of whom 18% currently subscribe. Market development: sell to offices, after 26 businesses ordered through the web shop last year without any marketing. Product development: canned cold brew for existing customers, which would need a co-packer and an estimated £60,000 before the first sale. Diversification: open a Kettlewhin café, at an estimated £250,000 fit-out cost with no retail experience in the team. For each it notes the evidence and the cost of a first test. The subscription push is cheap but the team judges the ceiling low; the office route already has paying customers and needs only a trade price list, a sample box and outreach to about 150 local firms over eight weeks. Cold brew goes on hold pending a survey of 400 customers, and the café is dropped for now. The roaster starts with office sales and sets itself a target of 30 office accounts before the next review.
Limits and criticisms
Two yes-or-no axes hide degrees of newness. An adjacent segment in the same country and a market on another continent both count as a new market, and a pack-size change and a wholly new product both count as a new product. Teams that use the matrix well place options along each axis rather than in a box, so that a near neighbour is not treated with the same caution as a leap.
The risk is implied rather than measured. The matrix assumes that more newness means more risk, which is a useful rule of thumb but not a calculation. A small, cheap diversification test can carry less money at risk than a large penetration campaign in a crowded market. The real comparison is between the evidence behind each option and the cost of being wrong, and the grid only prompts that comparison.
The matrix lists options but says nothing about attractiveness, capability or sequence. It cannot tell a large market from a small one, a move the business can deliver from one it cannot, or which move should come first. Market sizing, competitive analysis and a capability review must supply those answers. It also frames growth as choosing what to sell and to whom, so changes to pricing, business model or retention mechanics can be squeezed into the penetration cell where they are easy to overlook.
Finally, the cells are defined relative to one business. A product that is new to the firm may be familiar to the market, and a market that is new to the firm may be crowded with experienced rivals. Newness to the business is not the same as novelty in the market, and the plan should check both.
Related frameworks
The matrix decides direction; other frameworks fill in the detail. The STP framework is the next step for a market development move, because it decides which of the new segments to target and how to position for them. A SWOT analysis shows which strengths would travel into a new cell and which weaknesses would be exposed there. Before entering a market, the five forces model reads how much pressure buyers, suppliers, rivals, entrants and substitutes put on margin, and for a new country the PESTEL framework scans the regulation and conditions the move would face.
Common mistakes
- Leaving cells empty. A grid with only one filled cell is a justification for a decision already made, not a comparison of options.
- Defining the current market loosely. If "the market" means everyone, every move looks like penetration and the real unknowns stay hidden.
- Treating newness as yes or no. A neighbouring segment and a new continent need different amounts of research, budget and patience.
- Choosing diversification for its excitement. The most novel option should win on evidence, not on how it sounds in a planning meeting.
- Ignoring penetration because it is dull. Retention, frequency and cross-selling often have the cheapest evidence available, and should be tested before larger leaps.
- Stopping at the grid. An option is not a plan until it has been sized, targeted and turned into a launch with an owner.
